Tracking themes through fear & greed.

I spent my first hours on this assignment studying every playbook on the Alva explore page. My initial idea was to build something around my thesis research — tracking how master investors like Buffett and Dalio construct public narratives and how those narratives diverge from their actual portfolios. But someone had already done it. And more honestly — for a working investor making daily decisions, narrative archaeology is fascinating but not immediately useful.

So I went back to what I know. Five years in asset management taught me that what investors want most isn't more analysis — it's a quick read on the room. The concept I kept returning to was fear and greed: a framework every investor already understands, with real transmission power. When I'm investing myself, this is the kind of indicator I'd actually check before doing anything else.

The other thing I noticed: the best playbooks on Alva are deeply personal — one person's strategy, one person's edge. That's the community's strength. But the Content Leader role isn't about being another individual voice. It's about building infrastructure that the community can stand on. So I chose a topic that is foundational rather than opinionated. A Fear & Greed Index doesn't say "buy this." It gives everyone a shared baseline. Individual creators can remix it, layer their own thesis on top, compare their theme's temperature against the market. The official account provides the radar; the community provides the interpretation.

I hope you enjoy exploring it as much as I enjoyed building it.

— Xiaohan

Every theme has a sentiment cycle.

This template provides a standardized framework to quantify where any investment theme sits in its cycle, using a composite Fear & Greed score (0–100) built from measurable sub-indicators.

The template operates on two layers:

Layer 1 — Market Baseline. Broad US equity risk appetite, benchmarked to the S&P 500. Six sub-indicators, equal-weighted (17/17/16/17/17/16), refreshed daily.

Layer 2 — Theme Module. Theme-specific sentiment from a customizable basket. Default lookback: 126 trading days; exception: Realized Vol uses a fixed 252-day window.

Scoring: Each raw value → percentile rank within trailing window → mapped to 0–100. Inverse indicators flipped so 100 always means greed.

The divergence between layers is the actionable signal. Market greed + theme greed = beta-driven rally. Market neutral + theme greed = theme-specific overheating. Market fear + theme resilience = strong conviction.

A live playbook, two layers deep.

The 1-month delta of +39.4 shows the market moved from Fear (~35) to Greed (74) in a single month. AI trails by 9 points — but this gap is likely overstated: News and Social Sentiment (33% combined weight) are still accumulating. AI is the most discussed theme on social media right now; once live, those indicators will likely push the AI composite toward 70–75.

Sub-indicator snapshot reveals structural tension:

Prices are rising strongly (92) but the ride is bumpy (Vol 19). Earnings revisions are flat (50). This divergence is exactly the insight a composite score alone would miss.

Sub-segment leadership reveals intra-theme rotation:

• Memory
  100
• Compute
  92
• Power
  56
• Networking
  31

The headline 65 masks the fact that Memory is at extreme greed while Networking approaches fear. The AI theme hasn't cooled — it has fragmented.

Design decisions.

Why two layers? Theme sentiment can't be interpreted in isolation. Market 80 + theme 80 = beta. Market 50 + theme 90 = overheating. Market 25 + theme 60 = conviction. The two-layer structure makes these distinctions visible.

Why these indicators? Three signal types: price/technical (what the market is doing), fundamental (whether earnings confirm it), and narrative (what people are saying). This avoids the trap of pure price indicators missing narrative shifts, and pure sentiment indicators missing fundamental reality.

What I left out: Qualitative bull/bear cases (this quantifies sentiment, not thesis). Individual stock picks (the basket is input, not output). Valuation models (fair value ≠ sentiment). Put/Call ratio (requires Pro tier). These are deliberate boundaries — the framework tells you where you are in the cycle, not what to do about it.

Iteration history: v1.0 → v3.0.0. Three calibration fixes shipped during the build: gauge color palette, Flow z-score normalization (replaced binary percentile-rank with log-ratio logistic squash), and Vol lookback widened to 252d. Plus the sub-segment leadership widget — splitting the basket into Compute, Memory, Power, Networking by 20-day momentum.

What's next. Direction-adjusted volatility (high vol + up ≠ high vol + down). Full sub-segment composites once baskets have enough constituents for stable ranks. Sentiment backfill on Pro tier. Future indicator candidates: Polymarket odds, congressional trading signals, earnings call tone scoring, options skew.